Precision-Oriented Active Selection for Interactive Image Retrieval.
Résumé
Active learning methods have been considered with an increased interest in the content-based image retrieval (CBIR) community. Those methods used to be based on classical classification problems, and do not deal with the particular characteristics of the CBIR. One of those characteristics is the criteria to optimize, for instance the er- ror of generalization for classification, which is not the most adapted to CBIR context. Thus, we introduce in this paper an active selection which chooses the image the user should label such as the Mean Av- erage Precision is increased. The method is smartly combined with previous propositions, and lead to a fast and efficient active learning scheme. Experiments on a large database have carried out in order to compare our approach to several other methods.
Domaines
Machine Learning [stat.ML]
Origine : Fichiers éditeurs autorisés sur une archive ouverte
Loading...